Clinical immunotherapy trials are attempting to replicate cellular immune responses that can be curative in experimental tumor models. Though NK and T cells can be involved in murine antitumor responses, MHC-restricted tumor-specific T cells are more effective than NK cells in most models. Tumor-specific, MHC-restricted human T cells can be reproducibly demonstrated in vitro for many melanoma patients, however, identifying and manipulating this response has remained difficult for patients with most other malignancies. This may reflect defects in tumor antigen presentation and/or in T cell recognition. In contrast, nonspecific activation of T and NK cells can be induced with recombinant cytokines (such as IL-2) for most cancer patients. Antitumor effects are seen in some patients, but so is substantial toxicity to normal tissues. It is our hypothesis that the efficacy of current cellular immunotherapy regimens is limited more by the nonspecific reactivity of activated T and NK lymphocytes, than by the level of lymphocyte activation that can be achieved. Thus, we propose to use recombinant DNA technologies to express on patients' T and NK cells "artificial receptors" able to selectively recognize autologous tumor and trigger cellular immune reactivity. These tumor-specific antigen receptors have extracellular ligand binding domains composed of tumor-specific monoclonal antibodies, coupled to the transmembrane and cytoplasmic domains of receptor-associated signal transducing elements (the T cell receptor zeta chain, Fc receptor gamma chain, or the CD2 molecule). The goals of this approach will be establishment of a model system to characterize critical steps in lymphocyte recognition of, and activation by tumor cells, and to develop a clinically applicable system that could confer, through gene transfer, tumor-specific recognition capabilities to a patient's own NL and T cells. This approach would initially be tested in vitro, but could be developed in vivo in the form of a therapeutic gene transfer "vaccine." This proposal describes four years of laboratory research evaluating the ability of human lymphocytes to express functional tumor-reactive artificial receptors. As the production and screening of genetically engineered immunoglobulin (Ig) molecules is advancing, we anticipate other researchers will identify Ig molecules with even greater tumor specificity than those currently available. The present proposal seeks to develop a strategy whereby these "tumor specific" Ig molecules can be utilized as a component of efficient "artificial receptors" that could enable a significant cohort of a patient's T or NK cells to mediate specific in vivo recognition and destruction of autologous tumor.